amazon highlights: TOC Inventory Management / Namkee Chung / 2013

 TOC Inventory Management: A solution for shortage and excess dilemma / Namkee Chung / S.Korea / 2013

Generally speaking, inventory asset takes around 30% of total assets for a manufacturing firm and almost 90% for a distribution company. Note that there is no need of inventory management in case of bulk purchasing when a price raise is expected for certain materials or parts; there are fewer risks of maintaining excessive stocks as long as it is certain to gain the more profits with the more stocks. Also note that work in process or WIP is better to be dealt with operation management instead of inventory management since WIP inventory may vary depending on production planning method, production capacity of each process, and changeover system.

The core competency factor for distribution industry is a systematic control between sales and supply of goods. After all, it is impossible to avoid stock-out and overstock with the purchasing planning system based on the above demand forecast. Another method is needed that can function regardless of sales forecast. As such, there are great concerns about stock-out and overstock due to variation of demand when inventory management is solely at hands of a production planner to determine production volumes and timing. The key issues to be dealt with inventory management are rather decision-making issues as following than the data collection and application; - Selection of items for base stock - Location of inventory - Determining order quantity and timing - Decision for optimal inventory level

Firstly, to decide items for base stock, choose ones whose demand is large enough.

Secondly, location of inventory refers to a physical place for stored goods.

Thirdly, order quantity and timing strongly affect stock-out and overstock occurrences,

Lastly, an optimal level of inventory can be determined by ordering time and quantity as well.

Since an official record of 2-3% of stock-out rate would actually represent over 20% off-the-record rate, we need to keep the rate near zero. What would be the solution to keep the minimum stock while securing the maximum sales opportunities? It is only possible when the inventory control system is built up independently from demand forecast. improving sales forecast accuracy is not that easy. In fact, there is no ‘accurate demand forecast’ in real world, and that is the most prominent problem of inventory management. when the supply lead time is longer than the customer lead time, in order to prepare the required stocks in advance, there is no other choices but to place pre-orders by predicting quantity and timing. This is one of inventory control methods depending on demand forecast. The forecast-dependent pre-order system is frequently used also for goods in distribution. because each function of the supply chain seeks only a local optimization, it makes the supply lead time longer and the longer lead time brings the more dependency on sales forecast.

Inventory management should be implemented in a way to reduce the supply lead time rather than improving forecast accuracy. The inventory turns ratio is cost of goods sold (or net sales) divided by average inventory assets. despite the improved inventory turns, the number of unsold items would go up due to stock-out, and the sales would go down. Like previously mentioned, inventory should be controlled up to the point where it does not create any stock-out even during the process of improving inventory ratio. Even though we can fix the ratio by item, it could just be a target, and it does not imply any managerial advice. The inventory turns ratio itself does not provide any insight on how to decide an optimal inventory level nor how to manage it effectively. the suggested solutions are to get rid of the three main concerns – the reasons why operating a large inventory. - Make replenishment period shorter, - Improve demand forecast accuracy, - Enhance supply reliability

Two-bin system is to fix an order point and order quantity to facilitate the administration. This system is mainly useful for low value items, standardized goods with a shorter lead time, and office supplies.

Continuous Replenishment Planning (CRP) does not refer to a replenishment system itself but rather an operating program for it. It is a procedure or a scheme of business process of controlling inventory for goods for sale, indicating a method of replenishment for a certain location where goods are sold or consumed. Therefore, in order to implement CRP, it is a prerequisite to determine specific order quantity and time. In order to expect those benefits from CRP, a proper decision making process should be established so that it can make the best use of information on current supply, narrow variations in lead time, and cope with fluctuating demand and business environment.

Vendor Managed Inventory (VMI) is an arrangement where a supplier decides when and how much of customers’ stocks to be replenished. In that case, the vendor receives the payment only for those which have been consumed by the mother company. The VMI method can also be applied when a manufacturing company is managing inventory for a distributor. The seller supplies goods in accordance with the consumer demand except when there are sufficient stocks in the distributor’s warehouse.

Co-Managed Inventory (CMI) is similar to VMI in its structure of overall business process, but it is different in that suppliers and buyers are working together.

Supply Chain Management (SCM) is a business administration system to effectively operate the supply chain. SCM connects sales to manufacturing by regulating inventories in the supply chain. Related feedback from managers, supervisors, and staff is quite common, and can be summarized as follows; - Demand forecasting system is in a mess. - Inventory managers are negligent in monitoring stocks. - Purchasers buy too many for some items and too few for some others. - Information from sales/distribution is not sufficiently shared. - Customers change their minds too often. order cycle time and order quantity affect stock levels, as well as overall related cost for inventory controls; a frequent replenishment or a shorter order cycle time causes a heavier workload and more transportation cost, but on the contrary, expenses are likely to be reduced to maintain the inventory.

Replenishment can be done in four different models depending upon the decision factors for order time and quantity. That is, order time can be determined on either periodic or non-periodic basis, and order quantity can be decided on either fixed or non-fixed quantity basis.

Target Inventory System involves setting a target inventory and placing an order for what have been consumed during a given period of time. In other words, it is a replenishment method to make up the actual consumption for a cycle time. Placing orders at predetermined time intervals will facilitate and simplify daily duties. here are the summarized business processes; - Sales team to submit a replenishment request to plant (or distribution warehouse) to make up what have been actually sold. - Plant (or distribution warehouse) team to establish a production plan (or distribution plan) based on the request, while purchasing team to set a purchasing plan for materials/components in line with the actually produced volumes. - During this process, an individual target inventory (22 units of sales, 10 of plant, and 20 of purchasing team) enables each function to independently work on its planning. Such link in business flows between sales – distribution – manufacturing – purchasing is the core issue of Supply Chain Management (SCM) or Sales & Operation (S&OP), a major part of logistics management. can be summarized from the perspective of the Decoupled Pull System as below; - As sales team sells the goods, the stock level reduces in distribution warehouse as many as the sales quantity. Then, distribution warehouse team places a reorder to plant team (or a supplier) to make up the consumption. - The plant (or the supplier) makes a shipment out of its on-hand stock to the distribution warehouse, and adds up the same quantity to production planning. - The plant team places a reorder for the subsequent materials as much as it has consumed for production. To conclude, followings are the summarized features of Target Inventory System.

First, it is a decoupled pull-based system. Although order quantities match the actual sales, each functional operation can remain independent. This benefit reinforces the capacity to cope with demand variations by making full use of inventory functionality and by driving cross-functional cooperation.

Second, it streamlines the inventory controls as the order time can be fixed at a regular interval. There is no need to monitor stock levels all the time and no need to be concerned about missing an order time, so it reduces stock-out risks consequently.

Finally, what needs to be determined is the inventory target only. Order time can be easily fixed upon the corporate policy or upon the demand. To be specific, classify a set of products into A, B, and C group and fix the order cycle time as daily, weekly, and monthly each. Then, except under special circumstances, it is well enough to simply adjust the inventory target only although the demand or lead time may vary. It makes it easy to create an effective inventory management system.

It can be said that most of replenishment systems that are currently deployed in business are Reorder Point Systems. This method is still in use going through several modifications since Economic Order Quantity (EOQ) has been proposed in early 1900s. Indeed, a pull-based method is more appreciated than a push-based method in terms of effective demand controls in most of business management systems. The number of items in stock must increase in line with the more demand and the longer replenishment period. Replenishment Period needs be defined in detail; it consists of order cycle time, production lead time, and transportation lead time. First, order cycle time is the time between orders. If orders are placed on every Monday, then, its order cycle time is to be seven days. Second, production lead time can be defined as the total time required to produce an ordered item including order processing time, queue time, actual manufacturing time, and so forth. Finally, transportation lead time refers to the time from production finished to customer order delivered. Sometimes, the term ‘supply lead time’ can also be used, meaning the sum of production lead time and transportation lead time, that is, the time it takes from when you place an order until it arrives on your doorstep. Therefore, replenishment period can be said to be the sum of order cycle time and supply lead time. When the ordered items on Mondays arrive after five days, the replenishment period can be counted to be 12 days, 7 days of order cycle time plus 5 days of supply lead time. To sum up for a better understanding; - The amount of supply stock is determined by the demand during the supply lead time. The shop’s on-hand stock is determined by the demand during its order cycle time. Available stock, the sum of supply stock and on-hand stock, is determined by the demand during the replenishment period, i.e., supply lead time plus order cycle time. total available stock (= supply stock + on-hand stock) is proportional to the demand during replenishment period (= supply lead time + order cycle time).

However, generally speaking, people tend to be quite active in reflecting the demand rate but passive in reflecting replenishment period. This is why they put more efforts in forecasting demand. However, in fact, impacts of replenishment period are more critical having more indirect influences described as following; - In most cases, the longer replenishment period, the greater uncertainty, that is, the greater deviation of the period. - Hence, the longer replenishment period means the bigger demand variability. - Moreover, the longer replenishment period, the longer forecasting period of the demand. - Therefore, the longer forecasting period means the lower demand forecast accuracy. - Since target inventory consists of average demand for the replenishment period and safety stock to mitigate any variability, a doubled replenishment period should involve a doubled target inventory. Now, it is time to provide the target inventory formula. - Fix an expected demand for replenishment period. If demand rate is known, the demand forecast can be calculated by multiplying the demand rate by replenishment period. - Add safety stock to mitigate any uncertainty factor. This is determined by management policy considering a target customer service level.

Q = T – A Q: Order Quantity A: Available Stock (=On-Hand Stock + Open Orders – Actual Demand)

In order to lower stock levels and ensure sales by reducing stock-outs at the same time, it is necessary to shorten replenishment period. What are the main factors to make the order cycle time and supply lead time longer? One of the factors is the habitual practice of order processing in batch. Batch processing extends the order cycle time. Because many believe that a bulk purchasing lowers the purchasing price easing the order processing work, they take time to release an order until the orders are combined into a large batch. Besides, what makes a longer production lead time is production in large lots. People believe that decreasing set-up times while enlarging batch sizes is the best way to improve productivity and reduce production cost per unit, so ultimately it would create more profits. The reason why transportation lead time to be lengthened is also similar. Since making a big transportation batch can save transportation costs, people take time until the number of goods to be delivered becomes big enough before beginning transportation.

In order to make the replenishment period short, it is easier to reduce the order cycle time than supply lead time. Note that the shortened order cycle time can reduce on-hand stock but not supply stock. Supply stock can only be reduced with supply lead time decrease. Taking into account of only the demand factor, it is enough to run a system to set a shorter cycle time with a bigger demand or a longer cycle time with a smaller demand. In this way, it is determined by management policy based on experiences without meticulous calculations, and this is a better solution at a working level. For example, divide each set of products into A, B, and C group based on its sales quantity and unit price, then assign one week of order cycle time to A group, one month to B group, and three months to C group each. Simplicity works better.

Service level, in general, represents the percentage of satisfying customers’ needs. In inventory management, Service Level can be defined as the probability that goods are delivered as requested by customers. Safety Stock refers to an extra level of stock added to average demand to immediately enhance the service level. Safety stock is added only to on-hand stock and not to supply stock. Recall that the supply stock is determined by the demand during supply lead time, and on-hand stock is fixed by the demand for order cycle time. In this sense, safety stock is a buffer to mitigate demand variations during the order cycle time only, because there will be no more stock-out risk once the cycle time is finished thanks to newly arrived items. the decision factors of target inventory are demand rate (average demand per time unit), demand variability, replenishment period, and replenishment period variability. As a matter of fact, it is nearly impossible to attain a precise level of forecast. Even with heavy investment of time and money, forecasting accuracy is hardly improved. What we can expect to get is just some trend information on demand rate. if inevitably required, it is better to use short-term forecast rather than long-term one. Moreover, it is preferred to have forecasts by product group rather than by individual item. The deviation of the sum is normally smaller than the deviation of individual data, and this is what the Central Limit Theorem has generalized as a theorem. Although sales forecasting is a difficult task, an overall trend can be estimated to a relatively satisfactory level of accuracy.

Risk Pooling. It is a concept to effectively offset an overall uncertainty of the system by combining risk factors or uncertainties across locations into a single location. goals; - Make replenishment period shorter - Improve demand forecast accuracy - Enhance supply reliability –

Measures to shorten the replenishment period 1.     Make order cycle time shorter by placing orders for what have been actually sold on a previous day (or week, month, etc.). 2.     Make a small size batch to reduce supply lead time. 3.     Decrease or get rid of supply lead time by holding inventory.

Measures to improve demand forecasting accuracy 1.     Set a broad target area for demand forecasting. 2.     Integrate each regional inventory to a centralized distribution station for management. 3.     Make the best use of safety stock.

Measures to enhance supplier reliability 1.     Make the best use of safety stock. 2.     Reduce the number of emergency orders by placing orders for what have been actually sold on a previous day (or week, month, etc.).

Reorder Point System has its own value to be applied. The system has an advantage of being deployed when there is no substantial problems without sophisticated controls. The business processes can be much more simple. Supplies like raw materials/components are well suitable for Reorder Point system, those of which demand is not so big but that have a common usability. To be specific, the following is the process of inventory fine-tuning in preparation for seasonal variations. Estimate first the maximum sales during a high demand season. Check then available stocks on hand for this season. If you subtract the stocks on hand from the estimated maximum sales, you can get the target to be added before the high season comes. You need to raise the target level in advance up to this calculated point. Then, there will be no problem of supplying the demand as forecasted without having overstocks. On the other hand, you should not modify the target with a one-time sudden rise or drop that does not imply a trend. This can be settled by releasing special or extra orders. If the target inventory is changed with those one-time events, it will keep triggering stock-outs or overstocks afterwards. it is called Buffer Management that adjusts and controls the target inventory. Buffer Status (%) = On-Hand Stock level / Target Inventory These systems allow us to assess the priority not based on the smaller number of stock being more urgent but based on the ‘relatively’ smaller number of stock comparing with the target being more urgent.

“Instead of trying to be very precise in very uncertain situations (like using sophisticated forecasting techniques), TOC strives to build a robust design that is good enough. What complements good-enough planning is a very flexible and priority-driven execution control system that is capable of taking care of the exceptions.”

Buffer management allows to keep the optimal inventory by providing signals which orders need to get an urgent attention in response to demand variations including changes in demand rate, suppliers’ capacity, and supply lead time, etc. In case of penetration of the red-line, check first progress of open orders that have not yet been shipped to and call for an immediate supply as early as possible. On the contrary, in case of penetration of the green-line, postpone the shipment of open orders or do not place an additional purchase order. What needs to be done next is to verify how largely the buffer zone (the red or green) has been penetrated and modify the target level accordingly. This is what we call ‘Dynamic Buffer Management (DBM)’. The word ‘dynamic’ emphasizes here the meaning of active management and readjustment according to ‘situational changes upon time’. There are three types of circumstances in which the target level needs to be revised under Target Inventory System, described as follows;

1.     The target inventory has been set too low.

2.     The target inventory has been set too high.

3.     There are some special factors to be considered such as seasonality and promotions.

The size and duration of penetration are the two important parameters to adjust the target inventory. If the inventory penetrates ‘too much’ the red zone, the target level should be raised by one block size; the criteria ‘too much’ include either ‘too large’ size of penetration, ‘too long’ period of penetration, or ‘too big’ accumulated red zone penetration. After waiting until open orders arrive, that is, waiting for the replenishment time (this waiting time is so-called ‘Cooling Period’) passes by, you may start again to keep track of red zone penetration, if any. The reason why this cooling period is necessary is to check outcomes from the adjustment first before reviewing any further adjustment. If you add supplementary modification before the outcome is revealed, the adjustment process will never end that will make the target level go up endlessly.

If the inventory penetrates ‘too much’ the green zone, the target level should be lowered by one block size; the criteria ‘too much’ include either ‘too large’ size of penetration, ‘too long’ period of penetration, or ‘too big’ accumulated penetration Once these assessments on buffer penetration are done, the target inventory is to be adjusted downward, and you may begin monitoring the actual stock level again. Then, if the stock level falls nearly to the modified target, wait further until the replenishment time ends and now observe newly again the penetration situation of the green zone. Therefore, in this case, the cooling period is longer than the replenishment time. If there is no such cooling period in this process, the target level will almost endlessly go down.

In case where a sharp increase or sudden drop is expected due to the seasonality and promotion, there must be extra attention paid to this operation of buffer management. It can also be sought to temporarily suspend the buffer operation depending on situation. The graph shows the upward and downward temporary manipulations of the inventory level, ignoring the target inventory. When a drastic demand increase is foreseen, prepare in advance more stocks than the target, and stop the buffer management for a while to do so. When the demand is back to normal, you may resume operation of the system while keeping monitoring of the demand. In the same manner, when a rapid drop is expected in the demand, begin lowering the stock level regardless of the target. You need to stop the operation temporarily until the demand gets back to a stable level.

a replenishment period is the sum of order cycle time and supply lead time (= production lead time + transportation lead time). This supply lead time is sometimes very long comparing with order cycle time, and in such cases, target inventory becomes quite high relatively to on-hand stock. Moreover, the number of open orders increases, thus, the available stock (=on-hand stock + open orders – actual demand) increases accordingly because the quantity of open orders is proportional to the supply lead time

In these cases, the rules of determining buffer zones (green, yellow, red zones) must change. It may not be appropriate to keep the regular principle, i.e., setting each size of the zones to one third of the target inventory. If the same rule was applied to set the red zone to a third, stock-out risks from the on-hand stock could be very small regardless of its red sign. It means that the red zone signal does not work properly as a warning system because of enough open orders to be delivered soon. To make up for this kind of situation, it is desirable to fix the red zone size smaller than a third and the green bigger than a third. To sum up from the perspective of actual application, you have to make the order cycle time shorter and place orders more frequently, if it takes long time from the point of order to arrival, that is, if your product has a long supply lead time. While doing so, there are not much of stock-out risks because you have a plenty of open orders that will arrive soon despite the currently low level of on-hand stock. Moreover, if you set the red zone size smaller than a third of the target, the stock-out alarming system will become much more functional.

If you have sufficient available stocks despite a lower level of on-hand stock, potential risk of stock-outs will not be so significant. This is because there are ordered volumes to be soon delivered. In this case, rather than placing a new order, it is more effective to call for urgent supply of the open orders. In the case in which the available stock is not sufficient, you need to place a new order. If the on-hand stock is particularly not enough, you have to make it immediately. operating two separate buffer zones for two types of stocks can be very useful, especially when supply lead time is considerably longer than order cycle time. Hence, a good performance indicator on stock-out management should incorporate not only the quantity but also its duration period. Taking one step further, it should also involve financial loss caused by the stock-outs. In Theories of Constraints (TOC), this kind of measurement is called Throughput-Dollar-Days (TDD). It is a calculated value of a product’s throughput dollars multiplied by days of all the commitments not delivered.

Likewise, a good performance indicator for overstocks includes the length of stay. This value is calculated multiplying the overstock value (inventory dollars) by the number of extra days it is kept in the system. It is called Inventory-Dollar-Days (IDD) in TOC terminology. IDD will increase as the amount of overstock inventory goes up or as the days of stay gets longer. Remember that our objective is to manage the IDD to not to further increase or to make it reduced. TDD measures the financial value and delayed time of stock-outs, as well as loss coming from unfulfilled market demand. On the order hand, IDD measures the value of overstock that blocks flow of the system, and it is used as an indicator monitoring how fast each stock moves in developing SCM solutions, it is a valuable approach to set a target inventory and to distribute/produce according to actual sales quantity.

The forthcoming 4-step procedure is suggested to be followed,

1.     [Operate a centralized distribution station] Supplier (manufacturer or distributor) to establish a centralized distribution station and hold inventory for most of items in this location.

2.     [Regional warehouse to place orders on the basis of actual shipment] Each regional warehouse to fix a target inventory for each item and place an order for the same quantity as it shipped out on a previous day (or week) to the centralized distribution station.

3.    [Retail store to place orders on the basis of actual sales] Each retailer to fix a target level for each item and place an order for the same quantity as it sold on a previous day (or previous week).

4.     [Buffer management] Each of centralized distribution station, local warehouses, and retailers to keep track of demand variations and adjust the target level when necessary.

The objective of operating a centralized distribution station is to minimize forecasting errors. Taking into consideration of that, make a smaller target level towards the downstream to the retail level. To make it safe, keep the most of inventory at a center, a little less at a regional stock, and much less at retailers. Although the centralized warehouse will have an inventory rise, it is still less than the sum of  the decreased amount of inventory in each retail shop and regional warehouse. By rule of thumbs, it can cut the inventory nearly to 2/3 of the previous level. It doesn’t raise transportation cost although item composition and order cycle time has changed.

the starting point is to place most of inventory at supplier’s (or distributor’s) centralized distribution station so that retailers can place daily orders to the center (or regional warehouse) for the same quantity as the sales of previous day. Inventory is kept for common/flexible purpose Plant’s production planning becomes stabilized. Retailers’ inventory becomes reduced Retailers can have a larger space for displaying stands. Retail sales increase. Quality of fresh products is improved. New product launching is accelerated. New distribution channels are created. Overall Inventory level is reduced in distribution network It becomes less dependent on demand forecasting

in case of the materials of which supply lead time is longer that customer lead time, it is a must to secure the inventory in advance. you have to arrange the materials in advance to be available within the customer lead time. MTO (Make-to-Order) is a production system to start manufacturing of customized products only after a confirmed order for the products is received. MTO production method can be effectively operated by combining a replenishment system to MRP. Reserve inventory for the items that require to be held in stock and maintain the optimal inventory level through a replenishment system.